Unwanted and/or malicious network communications, such as spam, phishing, worm propagation, etc., hamper productivity and the use and enjoyment of computer and network resources by end users, burden affected networks with unauthorized and/or undesired traffic, and expose recipients to the risk of theft, fraud, etc. Current technologies for dealing with such unauthorized and/or unwanted communications focus on filtering messages at a point in the communication path close to the recipient system, such as anti-spam and/or other security software installed on the destination host and/or a local mail (or other messaging) server to which communications, to the destination host are directed for delivery to the destination host, e.g., by detecting messages sent by blacklisted senders and/or messages containing header or other content that indicates the message is suspicious. While this approach can be an effective way to protect an individual host or a group of destination hosts served by a local server, it does not address the network congestion problem and leaves vulnerable hosts not protected by destination (or destination mail or messaging server) based filtering software. In addition, in many cases an attacker causes each of one or more compromised hosts to send multiple instances of a malicious communication, often simultaneously or nearly so, magnifying the effect on the network. Therefore, there is a need for an effective way to intercept and take corrective action with respect to unauthorized, unwanted, and/or otherwise malicious electronic mail and/or other network communications that better protects the network and provides protection to destination hosts that are not protected by destination or destination mail or messaging server based filtering software.